Misrepresenting SFL Theory On Paradigmatic Relations

Fawcett (2010: 43n):

If you are not a systemic functional linguist, you may be asking at this point: "Why do systemic functional linguists give priority to paradigmatic relations between meanings rather than forms?" It is a good question, and it may be helpful to say briefly what my answer is. Ultimately, it is because generating a text involves making choices, and it is clearly the contrasts between alternative meanings between which we choose — rather than the contrasts between the forms. For example, if two outputs from the grammar display a contrast in form, as between that student and those students, the importance of the contrast is that the two forms express a contrast in meaning which the Performer wishes to communicate to the Addressee. In other words, the difference between 'singular' and 'plural' is ultimately a difference of meaning rather than form. (But there is, of course, no meaning without form.)

Blogger Comments:

This is misleading.  SFL theory gives priority to view from above, from meaning (semantics), in modelling the grammar (wording), but it nevertheless distinguishes between paradigmatic relations at the level of meaning (semantics) and paradigmatic relations at the level of wording (lexicogrammar).  Halliday & Matthiessen (2014: 49) explain:

Being a ‘functional grammar’ means that priority is given to the view ‘from above’; that is, grammar is seen as a resource for making meaning – it is a semanticky kind of grammar. But the focus of attention is still on the grammar itself.  Giving priority to the view ‘from above’ means that the organising principle adopted is that of system: the grammar is seen as a network of interrelated meaningful choices. In other words, the dominant axis is the paradigmatic one: the fundamental components of the grammar are sets of mutually defining contrastive features (for an early statement, see Halliday, 1966a). Explaining something consists not in stating how it is structured but in showing how it is related to other things: its pattern of systemic relationships, or agnateness …

On the other hand, SFL theory includes form in the paradigmatic relations at the level of wording (lexicogrammar) in the guise of the rank scale — clause, group/phrase, word, morpheme — each of which provides the entry condition for systems of paradigmatic relations between functions.

Misrepresenting Halliday On Form [2]

Fawcett (2010: 42-3):

The second key point is that, while it is incontestable that there are relations of contrast at the level of form, and while Halliday's concept of 'system' in "Categories" was, like that of Firth, a system of contrasts at the level of form, in a modern SF grammar the system networks model choices between meanings. And it is these that are seen as the generative base of the grammar. The result is that the purely formal contrasts in a language play no role in how the grammar operates in the generation of a sentence. … Thus choice between meanings is the key concept in a systemic functional grammar. However, the focus of this book is on the level of form, so I shall have very little more to say about the system networks.

Blogger Comments:

[1] This is misleading.  In Categories, Halliday (2006 [1961]: 39) uses 'form' in a different sense to that used by Fawcett:

The form is the organisation of the substance into meaningful events …

[2] This is misleading.  In "a modern Systemic Functional grammar", system networks model choices on all linguistic strata: meaning (semantics), wording (lexicogrammar) and sounding (phonology).  The lexicogrammatical networks model functional wording choices at each of the levels of form on the rank scale.  In the absence of grammatical metaphor, those functional choices at the level of wording agree (are congruent) with functional choices at the level of meaning.

[3] In "a modern Systemic Functional grammar", since form realises function, contrasts in form can realise significant contrasts in function, most notably in instances of grammatical metaphor, where what would congruently be realised by a clause is instead realised incongruently as a nominal group.  A major shortcoming of the Cardiff Grammar is its inability to systematically account for grammatical metaphor.

[4] To be clear, the focus of this book on Systemic Functional grammar is on neither system nor function.

What Fawcett Understands By 'Paradigmatic Relations'

Fawcett (2010: 42):

Paradigmatic relations are relations of contrast. There are two key points that must be made about them. Firstly, paradigmatic relations are unlike syntagmatic relations in that they exist only in the potential and never in an instance.  From the viewpoint of the text analyst, they express a contrast between (1) the meaning (and so the form) that was chosen for use in the text and (2) the one or more meanings (and so forms) that might have been chosen (but were not). In other words, paradigmatic relations exist only in the language that is used to produce a text-sentence — and not in the sentence itself.

Blogger Comments:

This continues Fawcett's confusion of the realisation relation between the paradigmatic axis (system) and the syntagmatic axis (structure) with the instantiation relation between potential and instance.  As previously explained, in Fawcett's model, schematised as Figure 4, paradigmatic system is equated with (meaning level) potential, and syntagmatic structure is equated with (form level) instance.

Fawcett On The Interplay Of Realisation And Instantiation In A Systemic Functional Grammar

Fawcett (2010: 41-2):

Let us return to Figure 4. Its significance is that it brings together, in a single diagram, two key pairs of concepts that correspond, broadly speaking, to two pairs of Saussurean concepts: meaning and form, and potential and instance. In a systemic functional grammar, meaning and form are related by the general relationship of realisation but, as we have seen, this relationship does not operate directly. Instead, it operates via the concept of instantiation. Instantiation occurs first at the level of 'meaning, when a traversal of the system network generates a selection expression of features, i.e. what Halliday has called an 'act of meaning' (Halliday 1993:4505). Then the realisation rules that specify the 'form potential' come into play and act upon the selection expression to realise it, and the final output from the grammar is the generation of a second 'instance', i.e., one unit that adds a layer of structure to the 'tree' representation of a text-sentence that is being built. 

Taken together, these concepts model the basic components of a systemic functional grammar, so that Figure 4 represents, at a fairly high level of abstraction, the main components of the model of language within which the alternative current theories of syntax in SFL can be set.

Blogger Comments:

This continues the discussion of Figure 4:

[1] This is misleading.  In a systemic functional grammar, stratal realisation doesn't "operate" and it doesn't do so "via the concept of instantiation"; this is Fawcett's misunderstanding only.  In a systemic functional grammar, realisation is an identifying relation between two levels of symbolic abstraction.  The notion of realisation "operating" derives from Fawcett's misunderstanding of the dimensions of SFL theory as interacting components, and the orientation of Fawcett's model to text generation by computer, rather than to human language itself.

Also in a systemic functional grammar, stratal realisation and instantiation are distinct dimensions, and, in terms of the theoretical architecture, form a matrix like the following:

[2] This is misleading.  This claim has not been supported by reasoned argument, merely asserted — and done so on the basis of theoretical misunderstandings.

[3] The terms 'first' and 'then' are misleading.  In a systemic functional grammar, there is no sequencing relation between two levels of symbolic abstraction.  The identifying relation is intensive, not circumstantial (temporal).  This again derives from Fawcett's misunderstanding of the dimensions of SFL theory as interacting components, and the orientation of Fawcett's model to text generation by computer, rather than to human language itself.

[4] This is misleading.  In a systemic functional grammar, realisation statements are located at the same level of symbolic abstraction as the network of features to be realised, not at a lower level.  Problematically, in Fawcett's model, a system network is realised by realisation statements.

[5] This is misleading.  In a systemic functional grammar, realisation statements apply to potential, not instances; the process of instantiation includes the activation of realisation statements.  Problematically, in Fawcett's model, a structure is an instance of realisation statements.  Problematically, in Fawcett's model, realisation statements specify an instance, rather than a realisation.  Problematically, in Fawcett's model, the realisation relation between paradigmatic axis (system) and the syntagmatic axis (structure) is confused with the instantiation relation between potential and instance.

[6] Given the theoretical (and logical) inconsistencies outlined above, it is very misleading to claim that Figure 4 genuinely represents a systemic functional grammatical model of language.

A General Model That Applies To All Systemic Functional Grammars

Fawcett (2010: 41):

At this point I must make it clear that Halliday sometimes writes in a way that implies a substantial change to the model represented in Figure 4, and the effects of this will be explored in Sections 4.6 and 4.7 of Chapter 4. I shall nonetheless argue there that Figure 4 does indeed represent a general model that applies to all systemic functional grammars.

Blogger Comments:

This continues the discussion of Figure 4:

[1] This is misleading.  The model represented in Figure 4 is entirely inconsistent with Halliday's model, as demonstrated in previous posts.

[2] This cannot be true, since the model represented in Figure 4 is internally inconsistent across both dimensions, stratification and instantiation, as demonstrated in previous posts, such as On 'The Main Components Of A Systemic Functional Grammar'.

The Prototypical Instance At The Level Of Form

Fawcett (2010: 41):

The prototypical instance at the level of form is a 'sentence' — and a sentence frequently consists of a single clause, e.g., I've been discussing that new student with Peter. Since the grammar is part of a fuller model for the generation of texts, we may also refer to the output as a text-sentence, and this has the value of reminding us that sentences do not occur singly, as formal linguists sometimes appear to assume, but within longer texts in which they themselves function as elements. (Note, however, that we can also treat a group of words such as that new student as an instance, exactly as is done in the little grammar in Appendix A.)

Blogger Comments:

Here Fawcett confuses the theoretical categories of 'instance' and 'unit'.  In SFL theory, any unit can be construed in terms of both potential and instance, and in terms of both system (paradigmatic axis) and structure (syntagmatic axis).

The Generation Of Instances Of Form

Fawcett (2010: 41):

In a fuller grammar the first unit to be generated would be a clause, and then one or more realisation rules would specify re-entry to the system network of meaning potential (as shown by the loop-back arrow on the left side of Figure 4), in order to generate one or more nominal groups (or even an embedded clause) to fill the relevant elements of the clause (as described in Fawcett, Tucker & Lin 1993). The 'tree structures' in the bottom right box in Figure 4 are labelled sufficiently richly to express the various functions that each element serves, and they are, it will be clear, the instances at the level of form.

Blogger Comments:

This continues the discussion of Figure 4:

[1] By definition, a realisation rule specifies a realisation (a lower level of symbolic abstraction), and so, not a re-entry to a system network at a higher level of abstraction (meaning) than form.

[2] The claim that 'tree structures are instances at the level of form' is merely a bare assertion, since no supporting argument is provided.  As Figure 4 illustrates, Fawcett incoherently regards tree structures as instances of realisation rules.  As the term 'realisation rule' makes plain, the relation between the rule and what it specifies is realisation, not instantiation.

The Rôle Of Fawcett's Realisation Rules

Fawcett (2010: 41):

In brief, we can say that the role of the realisation rules is to convert the selection expression of semantic features that is generated on a traversal of the network into a layer of the tree diagram representation of the sentence that is being built up. This concept is illustrated in Figure 2 of Appendix A, which has the potential to generate just eighteen different nominal groups. 

Blogger Comments:

[1] This blurs the theoretical distinction between the process of instantiation (the selection of features and the activation of realisation statements) during logogenesis ('the sentence that is being built up') with the realisation relation between the paradigmatic axis ('features') and the syntagmatic axis ('tree diagram').

[2] The illustration of the concept in Figure 2 of Appendix A is as follows:



It can be seen that this largely presents realisation rules as acting on individual features, rather than selection expressions, contrary to Fawcett's claims above, and that individual semantic features include the meanings of individual lexical items.

What Fawcett Means By The Instantiation Of Form

 Fawcett (2010: 40-1):

The key point is that, just as it is the activation of parts of the system network that specifies the output at the level of meaning (i.e., the selection expression), it is the activation of some of the realisation rules that specify the structural outputs from the grammar. It is the realisation rules — together with the 'potential structures', a concept that we shall meet in Section 9.[2].2 of Chapter 9 — that specify the structures, and that therefore constitute the 'form potential' in the grammar.

Blogger Comments:

This continues the discussion of Figure 4 (p36):


As even the term 'realisation rule' itself discloses, this misconstrues a realisation relation as an instantiation relation between potential and instance.

The Reasons Why Fawcett Prefers The Term 'Realisation Rule'

Fawcett (2010: 40n):

At various points in his writings, Halliday contrasts the systemic functional view of 'language as a resource' with the Chomskyan view of 'language as a set of rules'. Hence his strong preference for the term "realisation statement" over "realisation rule". Like many other systemic functional linguists, however, I take the view that, in defining the 'resource', we necessarily use a type of 'rule'. Thus a system network is itself a set of 'rules' about what features may be chosen under what conditions. This was first demonstrated in a fully explicit manner in the appendices to Hudson (1976), and similar 'rules' are found in the representation of the system network in a computer implementation in Prolog (as described in Fawcett, Tucker & Lin 1993). And realisation statements are even more obviously a type of 'rule'. In other words, while a systemic functional grammar does not have 'phrase structure rules' and 'transformational rules', it does have other types of rule. Here, then, we shall treat the terms "realisation rule" and "realisation statement" as interchangeable.

Blogger Comment:

[1]  The word 'rule' is problematic because encompasses two distinct types of modality: modulation  (obligation) and modalisation (usuality/probability).  As modulation, it also encompasses two distinct types of speech function: command and (modulated) statement; and the latter nullifies the distinction between 'rule' and 'statement'.

The term 'realisation statement', on the other hand, has the advantage of both specifying statement, rather than command, and encompassing probability (modalisation) as a property of system potential.

[2] This is an instance of the logical fallacy known as Argumentum Ad Populum, since it invokes the beliefs of (unspecified) others as support for the proposition.

[3] On the one hand, this is a bare assertion, unsupported by reasoned argument: the logical fallacy known as Ipse Dixit.  On the other hand, it is demonstrably false, since the notion of 'defining' does not entail the notion 'rule'.

[4] The use of thus here is misleading, since it gives the false impression that the statement that it begins follows logically from the preceding unsupported bare assertion.

[5] Since this is a bare assertion, unsupported by reasoned argument, the reference to Hudson (1976) constitutes an instance of the logical fallacy known as Appeal To Authority (Argumentum Ad Verecundum).

To be clear, a system network is organised on the basis of logical relations, such as:

• elaboration (delicacy)
• extension: alternation (disjunct options)
• extension: addition (conjunct options)
• enhancement: condition (entry conditions)

and to "read out" a traversal of a network is to produce statements of the type:

if X, then either Y or Z, and if both Z and A, then B or C.

For strict sense in which statements are a type of rule, see [1] above.

[6] Here Fawcett cites his own work as evidence in support of his own view.  This might be interpreted as the logical fallacies known as Appeal To Accomplishment and, on the basis of this critique, False Authority.

[7] This is another bare assertion, unsupported by reasoned argument.  For the strict sense in which realisation statements are a type of rule, see [1] above.

[8] On the one hand, this is another bare assertion, unsupported by reasoned argument.  On the other hand, it makes use of the logical fallacy known as Argument From Repetition (Argumentum Ad Nauseam).

[9] Here Fawcett, having purported to argue for 'rule' over 'statement', concludes by regarding the alternatives as interchangeable. 

On Instances Of Meaning As Inputs To Form Potential

Fawcett (2010: 40):

Figure 4 shows that at this level too — as we would logically expect — there is both a potential and an instance. The two levels of form and meaning are connected to each other, in a generative model of language, through the fact that the output from the level of meaning is the input to the level of form — more precisely, to the form potential. The form potential of a language consists principally of the realisation rules (or, as Halliday calls them, 'realisation statements').

Blogger Comments:

This continues the discussion of Figure 4 (p36):

[1] To be clear, on Fawcett's model, a structure is an instance of realisation rules.   As the term 'realisation rule' suggests, the relation here is realisation, not instantiation.  In this case, it is the realisation relation between the paradigmatic and syntagmatic axes (rather than the relation between potential and instance).

[2] As the terms 'output' and 'input' suggest, this is a model for text generation using computers, not a model of language as a human resource.  Moreover, the modular architecture of Fawcett's model is inconsistent with the dimensional architecture of SFL theory, where different levels represent different perspectives on language — different levels of symbolic abstraction — not different interacting components.

The Location(s) Of Phonology In Fawcett's Model

Fawcett (2010: 40n):

The Cardiff Grammar recognizes that it is only items that require expression in segmental phonology (which includes inherent word stress). One effect of this is that the two major aspects of phonology — intonation and segmental phonology — are treated as two separate components. They may look like one component when you view language 'from below', but if you look at intonation and segmental phonology 'from above', i.e., from the viewpoint of the meaning potential of the system networks, and if you then ask how meanings are realized in language, it becomes clear that the two are very different from each other: intonation realizing meanings directly, while segmental phonology does not. 

Blogger Comments:

[1] 'Inherent word stress' is not a feature of segmental phonology;  but see [2].

[2] The main theoretical disadvantage of treating intonation and segmental phonology as "two separate components" is that it omits rhythm from the model, since 'inherent word stress' is insufficient to account for the rich diversity of speech rhythms and the lexicogrammatical distinctions they realise.  The inclusion of rhythm is necessary for the modelling of intonation, since tone groups are realised by feet, and the ictus of each foot identifies the elements of potential tonic prominence, which in turn identifies the focus of New information.

[3] This misunderstands Halliday's 'trinocular perspective'.  It not possible to look at language 'from below', because there is no level of symbolic abstraction below language.

[4] To be clear, looking at phonology 'from above' means looking at it in terms of its function in various contexts (Halliday 2008: 141), as the expression of some content (Halliday & Matthiessen 1999: 504).

[5] To be clear, looking at 'how meanings are realised' — i.e. in terms of its various modes of expression — is looking at 'meaning from below' (Halliday 2008: 141).

[6] The claim here is:

• if you view meaning in terms of how it is realised,
• (then) it becomes clear that intonation realises meaning directly while segmental phonology does not.

This is not a reasoned argument, since no reasons are provided in support of the conclusion.  It is merely a bare assertion that has been dressed up to look like reasoning through the use of a conditional relation.  The advantages of such a model, to the theory as a whole, need to be both identified and supported by reasoned argument.

Modelling Content And Expression As The Same Level Of Symbolic Abstraction

Fawcett (2010: 39-40n):

There is a difference from Halliday's model in the way in which the term "form" is being used here. He uses "form" in a sense that includes (1) grammatical structures and items and (2) lexical items, but not intonation or punctuation. However, the Cardiff model of language integrates intonation and punctuation with syntax and lexis as the co-realizations of the meaning potential of the language, so that these too are regarded as types of 'form'. The effect is that intonation is not treated as 'below' the level of syntax and items, but as a parallel form of realization.

Blogger Comments:

[1] To be clear, it is grammatical units, not grammatical structures, that correspond to grammatical form in Halliday's model; that is, the compositional rank scale of clause, group/phrase, word and morpheme.  Grammatical structures, on the other hand, are function structures: i.e. function not form.

[2] To construe grammatical form and phonological/graphological form as the same level of symbolic abstraction is to construe content and expression as the same level of symbolic abstraction.  The distinction between content and expression is the major distinction of all semiotic systems.

[3] Here again Fawcett misleads by strategically confusing 'meaning' as a level of symbolic abstraction with 'meaning potential', the entire language conceived as a resource for making meaning (semogenesis).

What Fawcett Means By 'Form'

Fawcett (2010: 39):

We turn now to the level of form — and it is at this level that we require a theory of syntax. The term "form" is used here in a wider sense than that in "Categories" (or indeed any of Halliday's later writings) because it includes, as well as syntax and grammatical and lexical items, components for intonation or punctuation (depending on whether the medium is speech or writing). This is an approach to the concept of 'form' that looks at language 'from above', i.e., intonation and punctuation are here considered to be types of 'form' because, like syntax and items, they directly realize meanings.

Blogger Comments:

[1] To be clear, in Fawcett's model, content (syntax and grammatical and lexical items) and expression (intonation or punctuation) are located at the same level of symbolic abstraction.  The distinction between content and expression, as different levels of symbolic abstraction, is the fundamental distinction in semiotic systems.

[2] This misunderstands Halliday's 'trinocular' perspective.  On the SFL model, to look at language 'from above' means observing in terms of its function in various cultural contexts.  Halliday (2008: 141):

When we are observing and investigating language, or any other semiotic system, our vision is essentially trinocular. We observe the phenomenon we want to explore — say, the lexicogrammar of language — from three points of vantage. We observe it from above, in terms of its function in various contexts. We observe it from below, in terms of its various modes of expression. And thirdly, we observe it from its own level: from within, or from round about, according to whether we are focussing on the whole or some of its parts.

On the other hand, it is not possible to look at form 'from below' — i.e. as the content of some expression — because form is the lowest level of symbolic abstraction.  There is no lower level from which to look at form.  Halliday & Matthiessen (1999: 504):

A stratified semiotic defines three perspectives, which (following the most familiar metaphor) we refer to as ‘from above’, ‘from roundabout’, and ‘from below’: looking at a given stratum from above means treating it as the expression of some content, looking at it from below means treating it as the content of some expression, while looking at it from roundabout means treating it in the context of (i.e. in relation to other features of) its own stratum.

What Fawcett Means By 'Instances Of Meaning'

Fawcett (2010: 39):

Since there is a potential at the level of meaning, we should logically expect that there will also be instances at this level — and indeed there are. On each traversal of a system network, a set of semantic features is collec[t]ed, and the grammar then makes a copy of these, which is called a selection expression. There are two reasons for collecting the features as a set. The first is that they constitute the systemic description — and so, I would argue, the semantic description — of that unit in the text-sentence that is generated. The second is that the realization rules (to which we shall come in a moment) need to be able to refer to the whole set of the selected features, because many of the rules require two or more features to have been co-selected in order to 'fire', i.e., to be triggered into operation.

Blogger Comments:

This continues the discussion of Figure 4 (p36):

[1] On the SFL model, a traversal of a system network entails the selection of features, not the collection of them.  The notion of the grammar making a copy of collected features is not a model of humans engaged in the instantiation of texts, but a model of text generation by computer.

[2] This reason does not support the notion of collecting features, since the selecting of features, by itself, constitutes a "systemic description" of the instance.

[3] This "need" arises as a consequence of conceiving of a grammar as a flow chart between modules in which system networks and realisation rules are separated as two different levels of symbolic abstraction: system as 'semantic potential' and realisation rules as 'form potential'.  In SFL theory, the instantiation of systemic potential, at a given level of symbolic abstraction (stratum), involves both the selection of features and the activation of realisation statements.

For three internal inconsistencies in Fawcett's model see On 'The Main Components Of A Systemic Functional Grammar'.

Misrepresenting Halliday On Meaning Potential

Fawcett (2010: 38-9):

First, our model of language has, at the level of meaning, a component that specifies the meaning potential of the language — as Halliday has aptly named it (e.g., Halliday 1970:142). This is the core of a systemic functional grammar, and it consists of a vast system network of choices between meanings. In other words, the system networks model the language's potential at the level of meaning. Figure 1 in Appendix A introduces a simple system network for 'things', thus exemplifying the standard way of representing a system network in diagram form. 

Blogger Comments:

[1] This again misrepresents Halliday in order to give credence to Fawcett's model in which all system networks are located at the level of meaning.  For Halliday, 'meaning potential' is the entire language system, not merely the systems at the semantic level of symbolic abstraction (stratum).

[2] The system network of Figure 1 in Appendix A (below):

• confuses features (singular, plural) with what are specified by the synthesis of the most delicate features (water, bread etc.);
• confuses the semantics of things with the grammar of nouns (mass vs count, singular vs plural);
• confuses experiential 'thing' with both interpersonal deixis (near vs un-near) and textual cohesion (recoverable).

What Fawcett Means By 'Meaning'

Fawcett (2010: 38):

Now let us consider the term "meaning', as used in Figure 4. Throughout this chapter I have been careful to use to use the term "meaning" rather than "semantics" — even though I have happily used it elsewhere as the label for this level of language. Many systemic functional linguists (including Halliday in most of his writings) are understandably reluctant to use the term "semantics", because of the conceptual baggage that it brings with it from other disciplines and, within linguistics, from other theories of language. The types of 'meaning' that are covered in SFL by the system networks of TRANSITIVITY, MOOD, THEME and so one are much more comprehensive than the sense in which the term "semantics' is used by many linguists and philosophers. Nonetheless Halliday has till fairly recently allowed himself to use "semantic" (as a modifier) to refer to phenomena at this level of 'meaning'. And some systemic functional linguists — including Halliday himself in his important paper 'Text as semantic choice in social contexts" (1977/78) and myself — have regularly used the term "semantics" in the systemic functional sense of 'meaning potential'. We have done so because it is one way of expressing the theory's important claim that all of the different types of meaning covered by the system networks have to be included in any adequate theory of 'meaning', if only because the various sub-networks of TRANSITIVITY, MOOD, THEME and the others are partially interdependent on each other. SFL offers a particularly rich and powerful way to model the level of 'meaning' in language, and I have always felt it right to refer to this level of language by the term "semantics". Thus, in Figure 4.1 would be happy to replace "choice between meanings" by "semantic choices" and "meaning' features" by "semantic features".

Blogger Comments:

[1] This is misleading in a way that favours Fawcett's own position. It is Fawcett, not Halliday, who locates these systems at the level of meaning. In SFL, the system networks of TRANSITIVITY, MOOD, THEME are systems of wording (lexicogrammar), not meaning (semantics).  That is, they are posited as being of a lower level of symbolic abstraction than meaning — they realise meaning.

[2] Again, this is misleading in a way that favours Fawcett's own position of locating Halliday's grammatical systems at the level of meaning.  Halliday uses 'semantic' to refer to the stratum of meaning, with 'text' as the highest unit at that level of symbolic abstraction, and uses 'meaning potential' to refer to the theoretical construal of language as system.

[3] The argument here is:

• reason: because the more delicate metafunctional systems of the clause are partially interdependent on each other
• result: the meanings "covered" (realised) by the metafunctional systems of the clause have to be included in any adequate theory of 'meaning'.

It can be seen that this is a non-sequitur.  The latter does not follow from the former.  The partial interdependence of the more delicate metafunctional systems of clause is not itself a reason for the inclusion of all such systems in a theory of 'meaning' — any more than the "non-interdependence" would be.

[4] Rhetorically, this non-sequitur is presented as part of Fawcett's argument for locating Halliday's clause systems in semantics, rather than lexicogrammar.

[5] This is potentially misleading.  For Fawcett, modelling the level of 'meaning' is to interpret the metafunctional systems of the clause as semantic systems.  On the SFL model, on the other hand, modelling the level of meaning is modelling it as a higher level of symbolic abstraction (stratum) than all lexicogrammatical systems — not just those of the clause.  On the SFL model, the grammar not only realises the semantics, but makes possible the type of meaning that is only found in a tri-stratal semiotic system (i.e. language).

[6] Feeling that something is right is mere opinion, not reasoned argument.

What Fawcett Means By 'Lexicogrammar'

Fawcett (2010: 37-8):

In using the term "lexicogrammar" here, then, I am starting from the concept that a grammar is a 'model of language' (which is not the way that Halliday uses the term "grammar") and I am then incorporating into it, by prefixing it with "lexico", Halliday's important point that 'lexis' must be integrated with 'syntax' (or 'grammar') in any such model. But I have to point out that this is a hybrid term that does not correspond to Halliday's normal use of the term "lexicogrammar" — and, having made the point that the model must include lexis (and indeed intonation and punctuation), I shall normally use the shorter term "grammar" in the rest of the book, when referring to the concept of a model of the sentence-generating component of language.

Blogger Comments:

[1] To be clear, for Fawcett, a grammar is a model of language in which lexis is integrated with syntax — hence "lexicogrammar" — and which also includes intonation and punctuation.  When normally using the term 'grammar', however, he is only referring to the sentence-generating component.

[2] Halliday uses the term 'grammar' as shorthand for 'lexicogrammar' (wording), and locates its systems on a level of symbolic abstraction (stratum) between semantics (meaning) and phonology (sounding).  For Halliday, it is the grammar that construes (intellectually constructs) the semantics.  Halliday & Matthiessen (2014: 22, 10):

Grammar is the central processing unit of language, the powerhouse where meanings are created …

The clause is the central processing unit in the lexicogrammar — in the specific sense that it is in the clause that meanings of different kinds are mapped into an integrated grammatical structure.

[3] This is potentially misleading, since, for Halliday, syntax is not equivalent to grammar, but merely one aspect of it — modelled as the rank scale.

Misrepresenting Halliday On 'Form' [1]

Fawcett (2010: 37):

However, there is a problem about using the term "lexicogrammar" (or indeed "grammar") in a sense that includes the level of meaning. The problem is that Halliday has explored two different positions on the issue of what we might call 'levels of meaning', and when he uses the term "lexicogrammar" it is typically in a sense where it is equated with the level of form, such that this is in a relationship of realisation to the level of "semantics". (See Sections 4.6 to 4.9 of Chapter 4 for a full account of Halliday's two positions, and for the reasons why I think that his first position is greatly preferable.)

Blogger Comments:

[1] This is misleading in a way that supports Fawcett's stance.  Even before developing Systemic Functional Grammar, Halliday (2006 [1961]: 39) defines his theoretical levels as follows:

The primary levels are form, substance and context. The substance is the material of language: phonic (audible noises) or graphic (visible marks). The form is the organisation of the substance into meaningful events: meaning is a concept, and a technical term, of the theory. The context is the relation of the form to non-linguistic features of the situations in which language operates, and to linguistic features other than those of the item under attention: these being together “extratextual” features.

With the development and substantial elaboration of Systemic Functional Grammar, Halliday (1985) distinguishes between function and form on the lexicogrammatical stratum, with form modelled as the rank scale, from clause to morpheme, and with the choice of rank as the entry condition to the functions available at that rank.  In later editions, as if to answer misunderstandings like those of Fawcett, Halliday & Matthiessen (2014: 49) explain:

Being a ‘functional grammar’ means that priority is given to the view ‘from above’; that is, grammar is seen as a resource for making meaning — it is a ‘semanticky’ kind of grammar. But the focus of attention is still on the grammar itself.

[2] Fawcett's "full" account will be examined in detail in the critiques and clarifications of Sections 4.6 to 4.9 of Chapter 4.

What Fawcett Means By 'Grammar'

Fawcett (2010: 37):

At this point I should clarify the sense in which I am using the terms "grammar" and "meaning". Let us take "grammar" first, since it is used in the caption for Figure 4. In "Categories", "grammar" was the name of a subcomponent of the level of form, but here its meaning has been extended in two ways. The most obvious is that a grammar now includes a level of meaning as well as a level of form. Thus a 'grammar' is essentially a model of the sentence-generating component of a full model of language and its use. The second extension — which is less obvious — is that the term "grammar" is regularly used as a short form for "lexicogrammar".

Blogger Comments:

[1] Contrary to the implication, this "extension" of the meaning of 'grammar' to 'lexicogrammar' is not Fawcett's.  It appears in Categories of the Theory of Grammar (Halliday 1961), where lexis is conceived as most delicate grammar.  Halliday (2004 [1961]: 54):

The theoretical place of the move from grammar to lexis is therefore not a feature of rank but one of delicacy.  It is defined theoretically as the place where increase in delicacy yields no further systems;

[2] To be clear, Fawcett stratifies grammar into meaning and form, whereas Halliday locates meaning in semantics, and distinguishes between grammatical functions and grammatical forms (the rank scale).  Crucially, it is the distinction between semantics and lexicogrammar that allows Halliday to account for grammatical metaphor in his model.

Fawcett		Halliday
(lexico)grammar	meaning	meaning	semantics
		function	lexicogrammar
	form	form